1. Field of the Invention
This invention relates generally to testing of electronic components and more specifically to an improved system and method for the accelerated life testing of semiconductor devices in which a multi-purpose computer controlled driver system can accomplish the signal conditioning and testing of a wide variety of devices quickly and efficiently with a minimum of system setups and change-overs.
2. Description of the Prior Art
Accelerated life testing of semiconductor devices is a process which electrically ages these devices in their final and packaged form to help find defects which would result in premature failure. Although a variety of electrical conditioning and testing procedures may be used, most processes use heat as an accelerator by applying temperature stress to bring a defective semiconductor device to its failure point more quickly. Because of this use of heat, the commonly used term to describe such procedures is "burn-in" and the associated equipments are called "burn-in" systems. The required heat may be externally generated by placing the semiconductor devices in an oven, or by placing a heat source physically in contact with the semiconductor device package. The heat may also be self generated by electrically conditioning (biasing) the device to an extreme electrical condition.
Burn-in procedures were originally developed to prove that semiconductor devices would not fail early on in their operating life cycle. A key factor in their use to improve the reliability of semiconductor devices was the statistical analyses of the operating life of a given type device. These analyses usually showed a higher rate of failure during an initial "infant mortality" phase of operating life, a greatly reduced and stable rate of failure during the "normal operation" phase of operating life and, finally, typically after many years of operation, a gradually increasing rate of failure during the final "wear-out" phase of operating life. A further key factor was the statistical finding that "infant mortality" type failures could be caused to occur more quickly ("accelerated") through the use of heat and electrical over-stress. Thus failures of defective devices which might take months or years to occur under normal conditions could be caused to occur in just a few hours under burn-in conditions while non-defective devices were unaffected.
During the initial phases of the development of semiconductor devices, burn-in was widely used on products which had stringent reliability requirements to empirically demonstrate that the devices had survived the infant mortality phase of operation and were therefore reliable. Thus most semiconductor manufacturers have integrated burn-in into many of their intermediate manufacturing processes as well as a final test before shipment to a customer. As knowledge about the root causes of semiconductor device failure has increased and as the nature of the semiconductor manufacturing process has evolved to eliminate these root causes, the overall reliability of semiconductor devices has greatly improved. This improvement in reliability has occurred in a wide range of semiconductor devices including analog or linear devices and digital logic, memory or microprocessor devices and has resulted in changes in the way in which semiconductor manufacturers make use of their burn-in facilities. Instead of empirically confirming reliability by "burning-in" large quantities of the same device, manufacturers are using burn-in facilities to generate characterization and failure mode data on a wider variety of devices. By carefully controlling and recording the conditions which ultimately result in device failure and by performing thorough postmortem analyses to establish the precise root cause of each device failure, a statistical database can be created which allows continuing improvement in the design factors and the semiconductor processing steps used to make each device. These improvements in turn lead to still greater improvements in the overall quality and reliability in the semiconductor devices which have been analyzed and qualified through the "burn-in" process.
As a result of the above described increasing and crucial role of "burn-in" in the manufacture of reliable semiconductor devices, there exists a need for more sophisticated "burn-in" systems which make use of modular system elements which are readily reconfigured by computer control to accomplish the desired burn-in requirements and facilitate data collection and analysis for a wide variety of devices in a manner which is efficient and which requires a minimum of hardware reconfiguration.